Skip to content
Snippets Groups Projects
  1. Sep 22, 2023
  2. Aug 10, 2023
  3. Jul 06, 2023
  4. Jun 14, 2023
  5. Jun 12, 2023
  6. Jun 01, 2023
  7. Apr 03, 2023
  8. Mar 22, 2023
  9. Feb 27, 2023
  10. Feb 17, 2023
  11. Jan 27, 2023
  12. Jan 24, 2023
  13. Jan 09, 2023
  14. Jan 05, 2023
  15. Dec 02, 2022
  16. Nov 28, 2022
    • Christian Mauch's avatar
      Fix fstrings · 30240754
      Christian Mauch authored and Eric Müller's avatar Eric Müller committed
      Depends-On: topic:update_log4cxx, topic:spack_0.18.1
      In-Container: /containers/testing/c17764p47_2022-11-26_1.img
      Change-Id: I5106b6a29e8a140900bf5d5242ad90138f736b65
      30240754
  17. Nov 23, 2022
    • Philipp Spilger's avatar
      Fix install of all header files · 94b4b038
      Philipp Spilger authored
      * glob all header files in include folder opposed to white listing
        in order to also install detail folders and be future proof
      
      Change-Id: Ie7fedf2e910e5b77fb7a5218d0af46a5362f0fcd
      94b4b038
  18. Nov 22, 2022
  19. Nov 21, 2022
  20. Nov 14, 2022
  21. Oct 27, 2022
  22. Oct 21, 2022
    • Philipp Spilger's avatar
      Add API for parallel rng · 00c600b5
      Philipp Spilger authored
      * all vector types are supported, single vectors allow choice of
        which half of the rngs to use, half rows and vector rows use the
        whole rng, where for 16-bit rows the whole rng is queried twice
      
      Change-Id: I5a33dc19530b97a10e7d7b589bd6c3fd758e1a64
      00c600b5
    • Philipp Spilger's avatar
      Add vector_if selection operation · 296599aa
      Philipp Spilger authored
      * selects element-wise between two vectors based on a condition in
        a third
      * selection works if the element types are of the same size and the
        vectors have the same number of elements
      * all other element type combinations can be achieved by outside
        conversion of the mask type
      
      Change-Id: I5d41f9c44e3ad5aa486c9f658e8ba5d24180b82b
      296599aa
    • Philipp Spilger's avatar
      Add up and down conversions from 8 to 16 bit for VectorRow · f49b44e8
      Philipp Spilger authored
      * fractional saturating arithmetics are converted by using/setting
        the upper eight bits of the 16 bit value, thus preserving the sign
        and multiplicative behavior
      * module arithmetics are converted like static_cast using/setting
        the lower eight bits, thus preserving the value
      
      Change-Id: I993533e23cab4ad70fadc4826e2306c146bdef05
      f49b44e8
    • Philipp Spilger's avatar
      Add VectorRow{,{FracSat,Mod}{8,16}} · 8c0f1ead
      Philipp Spilger authored
      Change-Id: I1fea0f2ca7d8e1af6c9082798246f49e363a38c7
      8c0f1ead
    • Philipp Spilger's avatar
      Add VectorHalfRow{FracSat,Mod}{8,16} · 09571dce
      Philipp Spilger authored
      * 8-bit version is only typedef, 16-bit version contains two
        vectors, since they have half the elements each
      
      Change-Id: Id1dc3b39c590ebecf25a2839925ef1dd3439ad97
      09571dce
    • Philipp Spilger's avatar
      Add Vector{Mod,FracSat}{8,16} types wrapping __vector builtin · 39a6d92d
      Philipp Spilger authored
      * encapsulation of assembler instructions for arithmetics
      * provides conversion between different types of same element number
      * provides broadcasting construction and arithmetic
      
      Change-Id: I5732d2246da7d11544d75ea742d8d5f80569b8af
      39a6d92d
    • Philipp Spilger's avatar
      Add soft fractional number type · 93fe7fee
      Philipp Spilger authored
      * Keep header platform independent to allow prototyping on host
      
      Change-Id: If02f148a112b24bd8dd5d38f17b30941852eb673
      93fe7fee
  23. Oct 19, 2022
  24. Oct 17, 2022
  25. Sep 28, 2022
  26. Sep 21, 2022
  27. Sep 19, 2022